The present invention relates generally to an ultraviolet light-emitting device and a method of manufacturing a light-emitting device.
Group III nitride compound semiconductors such as, for instance, gallium nitride (GaN), aluminum nitride (AlN), indium nitride (InN) (hereinafter also referred to as a “Group III-nitride semiconductor” or “III-nitrides”) have been gaining attention as a material for semiconductor devices that emit green, blue or ultraviolet light. A light-emitting diode or a laser diode that emits blue light may be used for displays, for lighting and for high-density optical disk devices. A light-emitting device (which together with the acronym LED, when used herein, will for convenience also refer to both a light-emitting diode and laser diode unless otherwise specified) that emits ultraviolet radiation is expected to find applications in the field of ultraviolet curing, phototherapy, water and air purification, bio-detection, and germicidal treatment. The ultraviolet portion of the electromagnetic spectrum is often subdivided by wavelength into UVA (315-380 nm), UVB (280-315 nm) and UVC (<280 nm).
These LEDs are difficult to manufacture for a number of reasons. For example, defects arise from lattice and thermal mismatch between the group III-Nitride based semiconductor layers and a substrate such as sapphire, silicon carbide, or silicon on which they are constructed. In addition, impurities and tilt boundaries result in the formation of crystalline defects. These defects have been shown to reduce the efficiency and lifetime of LEDs and LDs fabricated from these materials. These defects have been observed for III-Nitride films grown hetero-epitaxially on the above mentioned substrates with typical dislocation densities ranging from 108 cm−2 to 1010 cm−2 for films grown via metal-organic chemical vapor deposition (MOCVD), molecular beam epitaxy (MBE), hydride vapor phase epitaxy (HVPE) and several other less common growth techniques. Reducing the dislocation density has accordingly become an important goal.
One way to reduce the dislocation density is based on the use of epitaxial lateral overgrowth (ELOG), which is a well-known technique in the prior art. With this method, the dislocation density can be reduced to about 105 cm−2 to 106 cm−2. This method, however, has been shown to be ineffective for the growth of aluminum-containing III-Nitride based semiconductors because of the tendency for the aluminum to stick to the masked material and disrupt the lateral overgrowth. Several variations of this approach have also been demonstrated including lateral growth (PENDEO) epitaxy, and facet controlled epitaxial lateral overgrowth (FACELO). All of these techniques suffer from the same limitation as the ELOG approach for aluminum containing III-Nitride materials.
Additionally, a technique called cantilever epitaxy involves growth from pillars that are defined through etching as opposed to, for example, masking.
Currently, several research groups are actively developing low-defect density AlN substrates to improve the power-lifetime performance of the deep UV LEDs. There are reports on a new air-bridge-assisted, high-temperature (1500° C.) lateral epitaxy approach to deposit 12-μm thick, high-quality AlN layers over SiC substrates as templates for the DUV LEDs. Pulsed lateral overgrowth (PLOG) of AlxGa1-xN has previously been demonstrated as an approach for depositing 15-20 μm thick AlxGa1-xN over basal plane sapphire substrates. Instead of the high temperature approach, a pulsed growth mode at 1150° C. was used to enhance Al-precursor mobilities over the growth surface. These pulsed, laterally overgrown (PLOG), AlxGa1-xN layers show a significantly reduced number of threading dislocations (˜107 cm−2) in the lateral-overgrowth regions, which enabled demonstration of optically-pumped lasing at 214 nm. In previous reports, the PLOG AlxGa1-xN was grown either from shallow (˜0.3 μm) trenched sapphire or from thin AlN etched templates (˜0.3 μm).
Yet another approach to decreasing defect density is a process referred to as pulsed lateral overgrowth (PLOG) wherein preformed layers are etched to islands. By controlling the flow rate of materials a layer is coalesced over the islands.
Several other approaches to dislocation reduction have been reported that do not involve selective area growth including inserting an interlayer between the substrate and the semiconductor layer to relieve strain, filtering dislocations by bending them into each other by controlling surface facet formation or by inserting a Group III-Nitride super-lattice layer as described in Applied Physics Letters, Jul. 22, 2002; Volume 81, Issue 4, pp. 604-606, between the buffer layer and the active layer.
Accordingly, several research groups at present are developing III-nitride deep ultraviolet (DUV) light emitting diodes (LEDs) for applications in air and water purification and bio-medical systems. Milli-watt power DUV LEDs on sapphire substrates with AlGaN multiple quantum well (MQW) active regions have been previously reported for the UVA, UVB and the UVC regions. The LED design used in the prior art comprises an AlN buffer layer deposited using pulsed atomic layer epitaxy (PALE), AlN/AlxGa1-xN, super-lattice layer between the buffer AlN and the n-contact AlGaN layer for controlling the thin-film stress and mitigating epilayer cracking; and a p-GaN/p-AlGaN hetero-junction contact layer for improved hole injection.
A majority of the current solutions for defect mitigation involve a superlattice. The superlattice is ultimately either a sacrificial layer or it is incorporated into the finished LED with no function. A sacrificial superlattice represents material which must be manufactured and scrapped thereby increasing manufacturing and material cost. If the superlattice is included in the finished LED as n-type doping in these superlattice is very difficult due to larger band gap. This insulating superlattice is detrimental to device performance. The thickness deviation of AlN and AlGaN leads to eventual cracking of the superlattice layer due to strain and lattice mismatch. Controlling the thickness of individual layers, quality of the epilayers and composition of AlGaN in superlattice layer is a major issue in growing high quality crack free thick AlGaN layers for high efficiency deep UVLEDs
In spite of the many developments and advancements there remains significant limitation for developing high power, reliable UVLEDs. Hence there is an ongoing desire for an LED and method for forming an LED with a low defect density.